Short answer: There is no single “best” cloud or developer platform, the category splits into hyperscale clouds (AWS, Google Cloud, Azure) built for scale and breadth of services, budget-friendly VPS providers (DigitalOcean, Akamai/Linode) built for simplicity and predictable cost, deployment platforms (Vercel, Netlify) built for shipping web apps fast, and a fast-moving layer of AI coding assistants (GitHub Copilot, Cursor) that now sit inside almost every developer’s workflow. The right choice depends on whether you need elastic scale, a predictable monthly bill, or speed of shipping, rarely all three at once.
This is a pricing and specification comparison built from each vendor’s official pricing page and documentation, checked in September 2026, not a hands-on trial of every platform.
-
Source
Official pricing pages and documentation for AWS, Google Cloud, Azure, DigitalOcean, Akamai Cloud, Vercel, Netlify, GitHub, GitLab, GitHub Copilot and Cursor, checked directly in September 2026.
-
Filter
Only published, current entry-tier and free-tier terms were used. Where a price is usage-based (per instance-hour, per GB, per credit), that metering model is stated rather than reduced to a single number.
-
Comparison basis
Entry price, free-tier terms and pricing model (flat, usage-based or credit-based), as published on each vendor’s own page. We have not run a hands-on trial of these platforms.
Cloud and developer-tool pricing changes frequently and is rarely a single flat number, model your actual usage with the vendor’s own calculator before budgeting.
What to check before choosing a cloud or developer tool
Flat monthly (VPS, most deployment platforms’ base tiers), usage/credit-based (Vercel, Netlify beyond free tier), or granular per-resource-hour billing (AWS, Google Cloud, Azure). This is the single biggest source of “surprise bill” complaints in the category.
Bandwidth/egress allowances and where the nearest region is to your users, egress is where hyperscaler bills quietly grow, while VPS providers tend to bundle generous transfer allowances into the base price.
Whether you can resize a running instance or autoscale a service, or whether you need to manually provision a bigger machine and migrate, this is where hyperscalers structurally outperform single-VPS setups.
Published SLA (if any), multi-region/multi-zone options, and what actually happens, credits, not refunds in most cases, if the SLA is breached.
Every major coding assistant beyond the cheapest tier uses metered credits for premium-model requests, rather than flat unlimited access.
What this category actually covers
“Cloud & Developer” is deliberately broad because the buying decisions in this space are connected. A team choosing where to run compute is also choosing how code gets deployed, how it’s monitored, and increasingly, what AI tooling writes and reviews that code. This pillar covers cloud platforms, VPS providers, APIs, developer tools, databases, DevOps and CI/CD, monitoring, deployment platforms, and AI developer tools including coding assistants.
One thing to state plainly before anything else: almost all cloud and developer infrastructure pricing is quoted in USD, even for UK buyers. AWS, Google Cloud, Azure, DigitalOcean, Vercel, Netlify, GitHub, GitLab and Cursor all list list-prices in dollars on their primary pricing pages. UK businesses pay in GBP via card conversion or, for enterprise accounts, negotiated regional pricing, but there is no separate “UK price list” for compute the way there sometimes is for consumer SaaS. Treat any GBP figure you see quoted elsewhere for these vendors as a converted estimate, not an official price.
Who needs this
- Startups and small dev teams choosing where to host a production app without a dedicated infrastructure engineer.
- Agencies and freelance developers who deploy client sites and need predictable, low-maintenance hosting.
- In-house engineering teams at SMBs evaluating whether to move off a single VPS onto a hyperscale cloud as traffic or compliance needs grow.
- Solo developers and technical founders picking a Git host, CI/CD pipeline, and, increasingly, an AI coding assistant as part of their daily toolchain.
- IT decision-makers consolidating vendor sprawl across compute, monitoring and deployment tools.
This category is not really for non-technical buyers browsing for a finished product; it’s for people who will configure, deploy to, and maintain what they buy.
Category types
Core cloud platforms, the hyperscalers with the broadest service catalogues (compute, storage, databases, AI/ML, networking) and the steepest learning curve: AWS, Google Cloud, Microsoft Azure.
VPS and simple cloud compute, single virtual servers or small clusters, priced per instance rather than per hundreds of services: DigitalOcean, Akamai Cloud (formerly Linode).
Deployment platforms and dev tooling, Git hosting, CI/CD and app deployment aimed at shipping code rather than managing raw infrastructure: GitHub, GitLab, Vercel, Netlify.
AI developer tools and coding assistants, in-editor and agentic tools that write, review and refactor code alongside developers: GitHub Copilot, Cursor, plus terminal-based agentic tools such as Anthropic’s Claude Code and Google’s Gemini CLI.
Adjacent layers not covered in depth here but worth naming: databases-as-a-service (Neon, Supabase, PlanetScale, MongoDB Atlas), and observability/monitoring (Grafana Cloud, Datadog). These deserve their own comparison pages rather than a paragraph each, flagged here so the category map is complete.
Key attributes that actually matter
Compute, what unit you’re buying (vCPU/RAM bundle, serverless function invocation, container) and how pricing scales with it. A hyperscaler prices compute by the second across dozens of instance families; a VPS provider prices it as a flat monthly plan per server size.
Network, bandwidth/egress allowances and where the nearest region is to your users. Egress (data leaving the cloud) is where hyperscaler bills quietly grow; VPS providers tend to bundle generous transfer allowances into the base price instead.
Storage, block storage (attached disks), object storage (S3-style), and database storage are priced and provisioned differently. Check whether backups and snapshots cost extra.
Scalability, can you resize a running instance, autoscale a service, or do you need to manually provision a bigger machine and migrate? This is where hyperscalers structurally outperform single-VPS setups.
Reliability, published SLA (if any), multi-region/multi-zone options, and what actually happens (credits, not refunds, in most cases) if it’s breached.
Pricing model, flat monthly (VPS, most deployment platforms’ base tiers), usage/credit-based (Vercel, Netlify, Grafana Cloud beyond free tier), or granular per-resource-hour billing (AWS, Google Cloud, Azure). This is the single biggest source of “surprise bill” complaints in the category.
Developer experience, CLI quality, documentation, first deploy time, and how much of the toolchain (CI/CD, previews, rollbacks) is included versus needing to be wired together separately.
Top products worth comparing
Amazon Web Services (AWS), The largest cloud platform by service breadth. EC2 (compute) is priced per instance-hour and varies by instance family and region; AWS’s on-demand pricing page describes the model as “per instance-hour consumed for each instance” rather than a flat rate, and all AWS accounts get 100 GB of free monthly data transfer out, aggregated across services (AWS EC2 On-Demand Pricing). Its free tier changed structurally for accounts created on or after 15 July 2025: instead of 12 months of always-eligible micro instances, new accounts get a $100 sign-up credit plus up to $100 in additional credits, usable over 6 months, with hard usage limits rather than the older uncapped-within-limits model (AWS EC2 Free Tier documentation). Best for teams that need the widest range of managed services and are willing to invest in the learning curve.
Google Cloud Platform, Offers $300 in free trial credits for new customers with no time limit until the account is activated for paid billing, plus an “Always Free” tier covering 20+ products at low, non-expiring monthly limits, including one e2-micro Compute Engine instance per month, 5 GB of standard Cloud Storage, and 2 million Cloud Run requests per month (Google Cloud Free). Strongest where a team is already using BigQuery, Cloud Run, or Google’s AI/ML stack.
Microsoft Azure, New customers get $200 in credit usable within 30 days, plus free monthly amounts of 20+ popular services for the first 12 months, and 65+ always-free services available indefinitely to any Azure customer (including Azure Functions’ 1 million free requests and Azure DevOps for up to 5 users) (Azure free account offer). The default choice for organisations already committed to Microsoft 365, Entra ID or .NET tooling.
DigitalOcean, VPS pricing starts at $4/month for a Basic Droplet (512 MiB RAM, 1 vCPU, 10 GiB SSD, 500 GiB transfer), rising to $24/month for 4 GiB RAM / 2 vCPUs / 80 GiB SSD (DigitalOcean Droplet pricing). From 1 January 2026 DigitalOcean moves to per-second billing with a minimum charge of 60 seconds or $0.01 (whichever is higher), a genuinely useful detail for anyone running short-lived instances. Best for teams that want cloud-like flexibility without hyperscaler pricing complexity.
Akamai Cloud (Linode), The cheapest shared-CPU compute plan is the Nanode 1 GB at $5/month (1 GB RAM, 1 vCPU, 25 GB storage, 1 TB transfer) (Akamai Cloud pricing, North America). Positioned similarly to DigitalOcean, simple, flat-priced VPS compute, but backed by Akamai’s global CDN network, which matters if edge latency is a priority.
Vercel, The Hobby (free) plan includes 1M edge requests, 100 GB of fast data transfer and 1M function invocations per month; the Pro plan is $20/month and includes $20 of usage credit plus higher allowances (10M edge requests, 1 TB transfer) before additional usage is billed (Vercel pricing). Built specifically around Next.js and frontend-framework deployment; less suited to arbitrary backend workloads.
Netlify, Free tier includes a 300-credit monthly limit; Personal is $9/month (1,000 credits), and Pro is $20/month with unlimited team members and credits starting at 3,000/month (Netlify pricing). Netlify’s credit-based model (replacing older flat “build minutes” pricing) is worth reading carefully before committing, since credit consumption varies by feature used.
GitHub, Free for individuals, with 2,000 GitHub Actions CI/CD minutes/month included (unlimited for public repos). Team is $4/user/month (promotional first-year rate) with 3,000 Actions minutes; Enterprise starts at $21/user/month (also a first-year promotional rate) with 50,000 minutes (GitHub pricing). Effectively the default Git host for most teams; the real decision is usually GitHub vs GitLab, not GitHub vs nothing.
GitLab, Free tier at $0/user/month; Premium at $29/user/month billed annually (currently bundled with $12 of included GitLab Duo AI credits per user/month as a promotional offer); Ultimate is custom-priced only (GitLab pricing). GitLab’s pull toward URIVIO’s audience is its built-in CI/CD and security scanning in one product, versus GitHub’s more modular ecosystem of separate tools and marketplace apps.
AI coding assistants, GitHub Copilot and Cursor, GitHub Copilot: Free tier gives 2,000 completions/month; Pro is $10/month/user with unlimited completions plus $15 of monthly usage credits for premium-model requests; Pro+ is $39/month ($70 in credits) and Max is $100/month ($200 in credits), with 1 AI credit equal to $0.01 (GitHub Copilot plans). Cursor: Hobby is free with limited agent requests; Pro is $20/month for an individual; Teams Standard is $40/user/month (Cursor pricing). Both use credit- or request-based metering rather than flat “unlimited” access once you move past entry-level use, which is the single most common source of confusion when teams budget for AI coding tools.
Comparisons worth reading before you commit
- AWS vs Google Cloud vs Azure, coming soon
- DigitalOcean vs Akamai Cloud (Linode), coming soon
- Vercel vs Netlify, coming soon
- GitHub Copilot vs Cursor, coming soon
Real-world use cases
A freelance developer deploying client sites. A VPS (DigitalOcean, Akamai) or a deployment platform (Vercel/Netlify free or Personal tier) covers most static-site and small-app client work without needing hyperscaler complexity. GitHub Free plus Copilot Pro rounds out a lean, low-cost personal stack.
An agency running many small client apps. GitHub or GitLab for source control and CI/CD, Vercel or Netlify for frontend deployment, and a VPS provider or Google Cloud/Azure for anything needing a persistent backend. The deciding factor is usually how many apps need to scale independently, more independent scaling needs push toward a hyperscaler.
A growing SaaS company outgrowing a single server. This is the classic AWS/Google Cloud/Azure decision point, when autoscaling, managed databases, multi-region failover, or compliance requirements (SOC 2, ISO 27001) start to matter more than raw monthly cost.
A developer or small team adopting AI-assisted coding. Start with GitHub Copilot Free or Cursor Hobby to test fit with your actual codebase and languages before paying for Pro/Pro+ tiers, the free tiers’ completion/request limits are generous enough to reveal whether the tool changes your workflow.
Pricing context: why “how much does this cost” rarely has a single answer
Outside of GitHub, GitLab, Copilot and Cursor, which publish flat per-user monthly prices, most of this category is priced on usage: per instance-hour, per GB transferred, per request, per credit consumed. That means the number on a pricing page is a starting point, not the bill you’ll actually receive. AWS, Google Cloud and Azure all publish official pricing calculators specifically because list prices can’t capture a real workload’s cost. Vercel’s and Netlify’s move to credit-based pricing has the same effect at a smaller scale, a $20/month Pro plan is a floor, not a ceiling, once usage exceeds included credits.
The practical implication for a UK buyer: budget in USD, add a currency-conversion buffer for card payments, and treat every “starting at $X/month” price in this category as the entry point for the lightest possible workload, not what a production app will actually cost.
How to choose
- Start from your actual workload, not the biggest brand. If you’re running one app on one server, a VPS is usually cheaper and simpler than a hyperscaler account you’ll only use a fraction of.
- Separate “where code runs” from “how code ships.” You can mix a VPS or hyperscaler for compute with GitHub/GitLab for source control and CI/CD, these aren’t mutually exclusive purchases.
- Read the free-tier terms, not just the headline number. AWS’s 2025 free-tier restructuring is a good example of why “free tier” details change and matter more than the marketing page suggests.
- Price AI coding tools by your team’s actual usage pattern. A $10/month Copilot Pro seat and a $20/month Cursor Pro seat solve overlapping but not identical problems, trial both against real work before standardising.
- Model your likely usage-based bill before committing, using the vendor’s own calculator where one exists (AWS, Google Cloud and Azure all provide one).
Common mistakes to avoid
- Assuming a flat “starting price” is the real monthly cost. It almost never is once usage grows past the free or entry tier.
- Picking a hyperscaler before you need one. The operational complexity (IAM, networking, service sprawl) is a real cost even when the dollar bill is low.
- Ignoring egress/bandwidth costs. Cloud bills are frequently dominated by data transfer out, not compute.
- Treating AI coding assistant credits as unlimited. Every major assistant beyond the cheapest tier uses metered credits for premium-model requests. Read the credit-to-request ratio before assuming “Pro” means “unlimited.”
- Comparing USD list prices as if they were GBP prices. They are not the same number, and no major vendor in this category publishes a separate UK price list for core compute or CI/CD products.
- Generous, genuinely usable free tiers exist across the category, Google Cloud’s Always Free tier, Azure’s 65+ always-free services, and GitHub’s 2,000 free Actions minutes/month all let a team evaluate a platform at zero cost.
- VPS pricing is simple and predictable, DigitalOcean from $4/month and Akamai Cloud from $5/month give a single-app team a flat, low-cost starting point without hyperscaler complexity.
- AWS, Google Cloud and Azure all publish official pricing calculators, giving buyers a way to model a real workload’s cost rather than relying on the headline entry price.
- Almost the entire category prices in USD with no separate UK price list, UK buyers pay via currency-converted card charges or negotiated enterprise contracts.
- AWS’s free tier changed structurally for accounts created after 15 July 2025, replacing 12 months of uncapped micro-instance usage with a capped, time-limited credit model.
- Credit- and usage-based pricing (Vercel, Netlify, AI coding assistants) means the advertised monthly price is a floor, not a ceiling, once usage exceeds included credits.
Is AWS, Google Cloud or Azure "the best" cloud platform?
None is universally best, each is best for teams already invested in its ecosystem (AWS: broadest service catalogue; Google Cloud: data/AI-heavy workloads; Azure: Microsoft-stack organisations). For a single app without complex infrastructure needs, a VPS is frequently the better starting point than any of the three.
Do I need a VPS or a hyperscale cloud account?
If you’re running one or two applications on one or two servers, a VPS (DigitalOcean, Akamai) is simpler and usually cheaper. Move to a hyperscaler when you need autoscaling, managed databases at scale, multi-region redundancy, or specific compliance certifications.
Is GitHub Copilot or Cursor better for AI-assisted coding?
They solve overlapping problems differently: Copilot integrates into existing editors (primarily VS Code and JetBrains IDEs) as an assistant; Cursor is a full editor built around AI-first workflows. Both offer free tiers, the practical answer is to trial both against your actual codebase before paying for either.
Will cloud and developer tool prices be quoted in GBP for UK customers?
Generally no. AWS, Google Cloud, Azure, DigitalOcean, Vercel, Netlify, GitHub, GitLab and Cursor all publish primary pricing in USD. UK buyers pay via currency-converted card charges or negotiated enterprise contracts; there is no standard published GBP price list for most of these products.
What changed with AWS's free tier in 2025?
For accounts created on or after 15 July 2025, AWS replaced the older model (12 months of eligible micro-instance usage within set hourly limits) with a credit-based free tier: a $100 sign-up credit plus up to $100 in additional credits, usable over 6 months, with hard caps rather than a rolling monthly allowance. Accounts created before that date keep the older terms. Always check current terms directly before relying on a “free” claim.
Is a free-tier cloud account enough to run a real production app?
Rarely, beyond early testing or a genuinely low-traffic side project. Free tiers are built to let you evaluate a platform, not to host production traffic. Budget for the paid tier from the start if the app is customer-facing.
Which cloud or developer tool should you pick?
- Freelance developer deploying client sites: a VPS (DigitalOcean, Akamai) or Vercel/Netlify's free or Personal tier, plus GitHub Free and Copilot Pro for a lean personal stack
- Agency running many small client apps: GitHub or GitLab for source control and CI/CD, Vercel or Netlify for frontend deployment, a VPS or hyperscaler only for apps needing a persistent backend
- Growing SaaS company outgrowing a single server: AWS, Google Cloud or Azure, once autoscaling, managed databases or compliance certifications start to matter more than raw monthly cost
- Developer adopting AI-assisted coding: start on GitHub Copilot Free or Cursor Hobby to test fit against your real codebase before paying for Pro or Pro+ tiers
There is no single “best” cloud or developer platform, the right choice depends on whether you need elastic scale, a predictable flat bill, or speed of shipping, and those three needs rarely point to the same vendor. Start from your actual workload rather than the biggest brand, model your likely usage-based bill with the vendor’s own calculator, and budget in USD with a currency-conversion buffer, since none of the major vendors here publish a separate UK price list for core compute or CI/CD.
Related categories
- Hosting, for managed WordPress, shared and dedicated hosting rather than raw cloud compute.
- AI, for AI writing, image, video and business-automation tools beyond coding assistants.
- Security, for VPN, endpoint protection and identity tools that sit alongside, not inside, cloud infrastructure spend.
Sources
- AWS EC2 On-Demand Pricing
- AWS Free Tier, EC2 usage documentation
- AWS Free Tier
- Google Cloud Free
- Microsoft Azure, free account pricing
- DigitalOcean Droplet pricing
- Akamai Cloud pricing
- Akamai Cloud pricing, North America
- Vercel pricing
- Netlify pricing
- GitHub pricing
- GitHub Copilot plans
- GitLab pricing
- Cursor pricing
- Grafana Cloud pricing
